Hi,
The performance of the pretrained model is not the same as reported in the main paper. Can you please tell me the exact steps you used to generate the images and evaluation script used to get the performance? I used the python script to evaluate the maps generated by MyTest.py in pytorch folder. I'm getting the performace as shown below:
<html xmlns:m="http://schemas.microsoft.com/office/2004/12/omml"
xmlns="http://www.w3.org/TR/REC-html40">
Hi, The performance of the pretrained model is not the same as reported in the main paper. Can you please tell me the exact steps you used to generate the images and evaluation script used to get the performance? I used the python script to evaluate the maps generated by MyTest.py in pytorch folder. I'm getting the performace as shown below: <html xmlns:m="http://schemas.microsoft.com/office/2004/12/omml" xmlns="http://www.w3.org/TR/REC-html40">
Dataset | | Sm (↑) | wFm(↑) | MAE(↓) -- | -- | -- | -- | -- CAMO | DGNet | 0.791 | 0.681 | 0.079 COD10K | DGNet | 0.776 | 0.603 | 0.046 NC4K | DGNet | 0.815 | 0.710 | 0.059